Last week, Efficient Frontier reported a 40% QoQ increase in Facebook CPCs, and this week TBG reported an approximately 30% drop in their U.S. campaigns. As Twitter chatter and the blogosphere debate the results, there are a few key differences in the numbers that have been missing from the discussion.

First, we reported on the actual CPCs that marketers paid while TBGs numbers reported trends in the suggested min and max CPCs on Facebook.  Therefore, the difference in percentage growth and decline  are an apples to oranges comparison.  Suggested CPCs are a function of Facebook’s algorithms and less a function of what advertisers are willing to pay for a click. In our opinion, CPCs are a better measure of the market because they reflect what people are ACTUALLY paying Facebook. Suggested CPCs are at best an inferential measure.

Second, we compared data from Q4 2010 to Q1 2011 while TBGs numbers trend from March 19th to April 11th, a period of about three weeks so it is a much shorter time frame in which to draw conclusions.

Third, the methodology of the analyses has not been mentioned. I would like to point out that for Efficient Frontier’s analysis, we examined a broad range of clients from a variety of sectors including retail, financial services, travel, and automotive.  Our index includes more than 10 enterprise class customers that have been Efficient Frontier clients on Facebook for more than 2 consecutive quarters. While we feel that our dataset is fairly robust and directionally correct, we are aware that there would be some variance in the numbers. This is typical of an early stage, rapidly growing channel where the marketplace is very dynamic. However, a +/-40% discrepancy is startling. Which brings us to the crux of the debate. Which number reports the correct trend?

Mathematically CPCs, click volume and Facebook’s revenue from Marketplace ads have a very simple relationship. Facebook’s marketplace revenue is the product of the CPCs that Facebook charges advertisers and the clicks that advertisers get .

If CPCs are truly down 30% the percentage of paid clicks in a quarter would have to increase by 30% for Fsacebook’s revenue to remain flat. Assuming that audience reach is relatively stable, the ad serving population would have to almost triple in a year in the U.S. for Facebook to record the same revenue as last year. Anecdotal evidence shows that such an increase in inventory is unlikely. As one commentor on AllFacebook points out, the roll out of the new profile in January would have actually decreased inventory as the “More Ads” link was removed.

This is contrary to all the trends and reports that predict that Facebook will double revenues in 2011. A look at the early days of search suggests that there will be some of both CPC increases and paid click increases going on. On the demand side, advertisers are rapidly leveraging the platform and are willing to pay more for their ads while on the supply side, Facebook is growing its audience reach so that there is an ever greater number of ads that Facebook can serve its ads. However, to record a doubling of revenues we believe that the bulk of the growth will come from the increase in CPCs at this stage.

In summary, there are 3 reasons why we think our numbers are directionally correct. First and foremost, we report on the actual CPC metric and not an inferential, indirect number that does not necessarily reflect the market. Second, we have a robust methodology that looks at longer term data. And finally, our numbers are consistent with the anecdotal evidence and reports about Facebook’s growth in 2011.

I hope this helps clear up the discussion.

Dr. Siddharth Shah
Sr. Director, Business Analytics